View Single Post
Old 08-26-2011, 09:48 AM   #5
Ted Craven
Grade 1
 
Ted Craven's Avatar
 
Join Date: Jun 2005
Location: Nanaimo, British Columbia, Canada
Posts: 9,238
Quote:
Originally Posted by mikesal57 View Post
I am not going into a debate on who's numbers are better.....
Dave's method was constructed using his numbers ..and it works for him
RDSS has Trackmaster.
Mike,

Of course it doesn't matter which data source is used - NewPace was sold demonstrating DRF SR's - what lower common denominator exists than that! I'm sure Dave does not intend that only HDW data makes it truly work.

Getting back to the specifics of this race, and the issue of ties - it appears you are trying to get the#5 and #7 into the mix and seeking to do so by extending the boundary of tied Low SR from 1 (or even 2) to 3 or 4 (the #5's low SR of 71 is 3 points distant from the tied 74 SR of the #2 and #4 and the #7 is 1 point even further away). Let's not forget, that with a repeatable (even semi-black-box) method, we are making our rules not just for one race scenario but for a thousand. I suspect that widening the tie area (even if done rigorously) greatly expands the fuzziness of the final contender definition process. Not that it couldn't be done, but in my opinion, it would have to be subject to other corollaries (e.g. the QESP points, or other RDSS factors for example) and tested broadly.

In refining the rules of NewPace, it might be good to always start by testing if ANY of the existing rules gets the job done in a given interesting scenario, like this race. The 'betting decisions' is one such category of rules, and I think one has to address exactly how many low odds horses (e.g. at or below 3-1) can be included in the betting contenders before the potential low return becomes too risky for the required outlay. I.e. seek first to try to eliminate one of the low price horses, and substitute with the best high price horse which was not previously a contender (aka 'price-only').

Sometimes this will get you the #5 horse, but if you apply the same rules 1000 times, we have to accept that a % of the time, we won't get the winner at all, or we will have concluded that the race was unbettable and that the correct decision was to pass.

Just my thoughts ...

Ted
__________________

R
DSS -
Racing Decision Support System™
Ted Craven is offline   Reply With Quote